Here's a design for a fresh water tank level sensor that uses LED indicators. Current draw of the chips is very low, on the order of 1 ma (if you use the CMOS versions), but the LED will draw up to 20 ma. I'll add some sophistication if there is enough interest in the community for a kit--a circuit that pulses the LED to reduce average current draw to about 4 milliamperes, low enough that the sensors can remain on all the time you're in the Airstream.
I took a look at discrete logic chips and it seems that 74LS (low power schottky) or 74HCT (CMOS) chips have a high enough input impedance that they can detect the connection due to imersion between the tank sensors. The early 70's Airstreams had 4 wires from the tank, a ground and 3 level sensors (I call them contacts--they are just screws through the side of the tank). The circuit in the graphic provides the following indications:
Green: Top contact in water/fluid
Yellow: Middle contact in fluid, top contact out of fluid
Yel/Red: Bottom contact in fluid, upper two out of fluid (this indication has two LEDs lit up)
Red: None of the upper 3 contacts is in the water/fluid (presumably the ground contact is always in the water)
I thought about getting four colors, but orange and blue are either not readily available or don't provide enough color discrimination. However, I wanted to show a definite difference between the lower contact in the water and the lower contact out of the water, to give 4 levels.
If there is sufficient interest, I'll lay out a small PC board, about 4x6", with three indicator channels--one for fresh water and two for holding tanks (the logic for the holding tanks is a little different, since they only have three wires--ground, half full, and full).
The cost of parts for what I've done herefor one channel is about $8--the most expensive piece would be a prototyping board from Radio Shack, about $5. The logic chips (74xx04, hex inverter, and 74xx08, quad 2-input nand gate) are like $0.40 each, plus LEDs, resistors and a small three-terminal voltage regulator. You can add a second channel (black tank) for another $2, just resistors and LEDs, since the two chips have enough gates for two channels. For everyone with a pre-1975 trailer, you ony need two channels. That's it.
The prototyping board is laid out on 0.1" centers, perfect for chip sockets and spacing the LEDs. You'd have to make a new bezel on your own. If you wanted to use the existing panel (mine has
12V light bulbs--you'd have to add another chip that has open-collector outputs to handle the voltage. It's only another $0.40, but it's another 10 wires on the board and then all thew wires to the bulbs) there is a way to do it by wiring the existing light bulb sockets into the board. The logic is the same.
For 1975 and later, you need three channels (fresh water, grey and black holding tanks). But there is a twist. The three sensor wires are shorted together and the meters that indicate level sense a total resistance, as each of the sensors (contacts) is immersed. I need to do a little experimenting to see if I can create a small system to duplicate that method method of using only two wires, since only two wires show up at the control panel.
A printed circuit board that would go together without any discrete wiring would add about $30 to the project. I'd need to hear from 20 or so members who were interested in order to make that happen. Don't let all the above ancient electronic stuff scare you away. The PC board would be simplicity itself, compared to taking care of rear end sag and polishing!
The resistors can't be specified without further experimenting. The pullup resistors worked fine at 56K, but my quick test of water impedence showed 50-200K with the probes in the water, so I'd be more confident if the pullups were about 350K (for CMOS--none needed for schottky). The series resistors for the LEDs are necessary to limit the current to 20 ma max, but there is a further complication. The different LED colors have different efficiencies in converting current to light. So you would want to balance the brightness by using appropriate resistors to modify the current. For example, even low efficiency green, yelow, and red LEDs have brightness specifications varying from around 5 millicandles to 700 mcd. So the series resistor would vary from about 250 ohms to 5000 ohms.
It'll be a the middle of September before I can refine this. Let me know if you're interested.